[02:19:06] I tried to login to beta labs, it failed trying to access en.wikipedia.$channel.wmflabs.org. < $channel, not beta. Known issue? [02:21:27] Special:UserLogin redirects to http://en.wikipedia.$channel.wmflabs.org/wiki/Special:CentralLogin/complete?token= [02:32:38] I actually am logged in, despite the $channel failure [03:10:54] spagewmf: that's my fault; fixing. [03:12:11] ori: thanks. I filed https://bugzilla.wikimedia.org/show_bug.cgi?id=63756 [03:12:26] many thanks! [09:40:55] my proxycommand stopped working :/ [09:41:00] just in time andrewbogott :) [09:41:13] matanya: ok, just a second... [09:43:30] matanya, would it help if I say that it's still working for me? [09:44:39] matanya: how long ago did it stop working? For example, last week I shut down the pmtpa bastions so you probably need to point at eqiad now... [09:48:39] matanya: did you lose your backscroll? [09:48:43] yes [09:48:57] If it just broke today then I'm not sure what's up. [09:48:58] my computer swapped died [09:49:10] i didn't try it since migration [09:49:14] If you haven't tried in a few days… maybe the issue is that the pmtpa bastions don't exist anymore? [09:49:21] What does your proxy command look like? [09:50:13] e.g. does it use an ip? Or are the bastions in your hosts file? [09:50:30] Host *.wmflabs.org [09:50:30] ProxyCommand ssh -a -W %h:%p bastion.wmflabs.org [09:50:42] hm… that should still work. [09:50:55] But you probably need to clear your fingerprint too… https://wikitech.wikimedia.org/wiki/Help:SSH_Fingerprints [09:50:58] for bastion I mean [09:50:59] i have [09:51:09] that was the first thing i checked [09:51:14] can you ssh to the bastion directly? [09:51:35] yes [09:51:46] without the command of course [09:52:48] that's the same as mine… approximately [09:52:55] What kind of a failure do you get? [09:53:05] You've tried with different instances I take it? [09:53:15] nothing, just pending for ever [09:53:21] i have [09:53:49] try puppet-testing-1.eqiad.wmflabs [09:54:11] matanya: What are you trying to connect too. [09:54:22] https://tools.wmflabs.org/paste/view/22c645e7 [09:54:22] to*? [09:54:43] do i need to specify eqiad ? [09:55:09] well of course: ssh: Could not resolve hostname puppet-testing-1.eqiad.wmflabs: Name or service not known [09:55:54] matanya: That WMF too, to that same box (rt-128) [09:56:09] what coren ? [09:56:38] yes, can't resolve [09:56:49] I think your proxy command name doesn't make sense. Mine looks like... [09:56:57] matanya: Wait, that ProxyCommand could never have worked -- *.wmflabs.org are public IPs [09:57:07] Host *.eqiad.wmflabs [09:57:08] ProxyCommand ssh -a -W %h:%p bastion-restricted-eqiad.wmflabs.org [09:57:16] and then I would ssh rt-128.eqiad.wmflabs [09:57:25] Well, 'Host *.wmflabs.' also works. [09:57:26] (Of course bastion-restricted won't work for you) [09:57:29] oops, i gave you the wrong proxy command :/ [09:58:10] ok [09:59:13] well, i broke something as it seems [09:59:18] not sure what [10:00:05] I'm pretty sure that the hostname that gets passed by the proxy command needs to be an actual host that the bastion can see. So that's probably what to look for. [10:00:11] (Just paged for lunch so we're about to vanish :) ) [10:00:28] (We're all in UTC+3 today!) [10:00:42] me too! :) so what should have been in ProxyCommand ? [10:00:48] * yuvipanda sets out LSD drops in the rain around Athens [10:00:53] ProxyCommand ssh -a -W %h:%p what ? [10:01:44] I'm not sure… the command as I pasted it before definitely works. [10:01:45] But, gotta go! [10:01:51] see ya [10:01:54] LSD kicking in [10:02:02] ori: you might know ? [10:02:30] oh, I had a similar issue [10:02:35] i think something is borked, yes [10:02:51] i had a bunch of ssh alias to the I<....>.eqiad.wmflabs hostnames and they all stopped working [10:03:06] i had to replace it with the "friendly" instance name in each case [10:03:39] * matanya wonders what [10:04:16] i can get to basion but proxing doesn't work for some reason [10:04:24] probably layer 8 issue [10:06:42] matanya, andrewbogott: http://p.defau.lt/?hPUWKFTmBUNMLlKIfzlKLw [10:07:05] hmm [10:07:11] and Coren ^ [10:08:21] ok, i fixed my issue [10:08:36] with the nice names solution [10:08:39] thanks ori [10:24:51] Anyone care to suggest why Faebot's webtools are not running? I seem to have migrated fine and /webservice restart/ appeared to run okay. Example http://tools.wmflabs.org/faebot/cgi-bin/TARDIS.py [10:41:29] Okay, I was helped to map my cgi-bin which has got it going again. [11:18:59] Hi! Danny_B What about *your* tools on TS? Do you have a Tool Labs account with a different name? [11:20:47] matanya: sorry to run off… you're up and running for now? [11:20:54] (I see that you updated docs but haven't read your change yet) [11:21:03] yes andrewbogott i'm ok [11:21:09] thanks [11:21:14] 'k [11:21:33] If you think there's a fundamental problem as well, I'm interested, but can't think about it today… best to log a bug [11:22:21] andrewbogott: not fundamental, just can't access instances with there I-number names [11:22:33] which isn't critical [11:22:53] most of my issue was layer 8 issue [11:23:17] ok [14:31:39] Coren, archive_userindex doesn't appear to be indexed very well [14:31:57] Coren, it takes 18 seconds to run a query on it. [14:32:23] Coren, SELECT COUNT(*) AS count FROM archive_userindex WHERE `ar_user` ='14836860'; [14:38:24] Coren, [14:38:25] +-------+ [14:38:25] | count | [14:38:25] +-------+ [14:38:25] | 344 | [14:38:25] +-------+ [14:38:30] 1 row in set (16.76 sec) [14:43:06] petan, I still stand by pinging people first. [14:43:18] Cyberpower678: I don't give a fuck [14:43:40] petan, like python? [14:43:42] :p [14:43:51] !python [14:43:51] There are multiple keys, refine your input: pythonguy, pythonwalkthrough, [14:43:55] :/ [14:44:02] WTF?? [14:44:08] Where did it go? [14:45:13] someone deleted it [14:45:33] !python is http://fieldbio.wikispaces.com/file/view/python.jpeg/114402171/python.jpeg <-- Python, waiting to eat all your resources, and you. [14:45:33] Key was added [14:46:24] petan, ^ I put it back. [14:46:55] silly kids removing important material [14:47:16] Scrollback tells me [14:47:17] !python del [14:47:17]

Successfully removed python [14:47:17] ok [14:47:28] Cyberpower678: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Tables_for_revision_or_logging_queries_involving_user_names_and_IDs [14:47:41] Cyberpower678: No word about archive. [14:50:00] scfc_de, it doesn't say why archive_userindex is slow when I use ar_user. The principle for all userindexes should be the same. [14:51:11] File a bug so it can be taken care of? [14:52:38] !log deployment-prep Adding Tobias Gritschacher to the project so he can look at udp2log / apache logs whenever needed :-] [14:52:41] Logged the message, Master [14:53:10] scfc_de, couldn't it be easily fixed by reindexing the table? [14:56:08] Cyberpower678: a) I don't know that; only Coren and Sean have access to the replica DB servers. b) I'm very certain that we don't want a solution that solves querying archive_userindex in enwiki_p once, but everywhere for eternity :-). [15:35:34] scfc_de: any idea why I might be getting: [15:35:35] -bash: /data/project/android-sdk-builds/android-sdk-linux/build-tools/19.0.3/aapt: No such file or directory [15:35:43] when the file clearly exists [15:35:45] and has x bit set? [15:35:46] ls -l tells me [15:35:52] -rwxr-xr-x 1 tools.android-sdk-builds tools.android-sdk-builds 1109734 Apr 10 15:19 /data/project/android-sdk-builds/android-sdk-linux/build-tools/19.0.3/aapt [15:37:36] scfc_de: puzzle! :D [15:43:24] scfc_de: weird strace https://dpaste.de/OQpU [15:52:13] yuvipanda: If I copy it to my home, the error remains, so it is in the file? [15:54:26] yuvipanda: "Unix's most misleading error message": http://stackoverflow.com/questions/5234088/execve-file-not-found-when-stracing-the-very-same-file [15:58:03] yuvipanda: Another oddity: "file aapt" => "dynamically linked (uses shared libs)", "ldd aapt" => "not a dynamic executable". Did you download that file or was it compiled on Tools? [16:13:16] scfc_de: I downloaded it [16:15:30] scfc_de: looking [16:15:33] scfc_de: no, it was downloaded [16:18:08] yuvipanda: Same error two years ago: http://pastebin.com/CUKCqGey [16:18:49] Don't know; maybe different architectures? [16:19:38] scfc_de: hmm, it is supposed to work for both [16:23:19] scfc_de: it's the android sdk, and for linux you are just supposed to download and run [16:29:23] yuvipanda: https://developer.android.com/sdk/installing/index.html says "If you are running a 64-bit distribution on your development machine, you need to install the ia32-libs package" (I'm pretty sure Labs is 64). But "aptitude install ia32-libs" brings up lots of conflicts with installed packages. [16:29:35] oh fuck, right. [16:29:51] I completely forgot about that [16:31:33] Do you remember the solution? :-) I don't think Android would require its developers to downgrade their systems. [16:31:34] scfc_de: we're on 12.04 right? [16:31:59] 12.04.4 LTS, Precise Pangolin [16:32:58] scfc_de: right. [16:33:16] What do you use in the app team? Or does every developer take care of his own system? [16:33:28] scfc_de: everyone is on OS X [16:33:30] * yuvipanda shames self [16:33:35] scfc_de: this is me attempting to setup CI [16:33:41] scfc_de: http://askubuntu.com/questions/300139/ia32-libs-broken-package-ubuntu-12-04-x64 has a solution [16:35:30] Oh, CI, just file a bug and let hashar figure it out :-). [16:35:52] scfc_de: hehe, we're at the part of the cycle where I am figuring it out for hashar :) [16:36:48] scfc_de: plus this might just end up being a cron on toollabs. [16:37:00] scfc_de: I already needed a local install of maven... [16:38:24] scfc_de: I tried it on another instance (android-build, mobile project) [16:38:29] sudo apt-get install libc6:i386 libncurses5:i386 libstdc++6:i386 libz1:i386 libsdl1.2debian:i386 [16:38:34] that seems to make it work [16:40:31] Before I screw up tools-login or -dev, I'll create a new instance for tests. Might take an hour or so for the initial Puppet run. [16:40:41] scfc_de: I have one (android-build instance) [16:40:54] scfc_de: I tested there, works fine, and they are on -login now :P [16:40:54] should've put them on -dev [16:41:03] scfc_de: but how do I puppetize these? [16:42:15] yuvipanda: Does android-build have role::tools::something? [16:42:23] Re puppetizing, no idea. [16:42:26] scfc_de: none at all, no. I just apt-geted them [16:44:11] It's the conflicts with the installed packages that makes this interesting :-). [16:44:56] http://projects.puppetlabs.com/issues/265 only has info about Puppet pinning architectures with Yum. [16:51:05] anyone know why login on beta labs is busted? https://bugzilla.wikimedia.org/show_bug.cgi?id=63780 (after logging in resulting page URL is bogus) [16:51:11] scfc_de: but if I install the individual packages there's no conflicts [16:54:45] bd808: heads up. ^^ people messing with Apache configs in beta labs recently? [16:55:30] chrismcmahon: It may be me yes. I'll look into it. [16:57:10] bd808: I can tell you that 19 hours ago beta was OK but about 8 hours ago it was not OK [16:57:58] chrismcmahon: Ok. I can tell you that about 14 hours ago I was messing with with beta deploy mechanics. [16:58:26] bd808: sounds suspicious :-) thanks for looking. [16:59:11] Login does seem to work but the redirect URL is borked as you said. [17:43:08] Coren: are you about? [17:45:35] Can somebody add me to the HHVM group on labs? [17:45:41] My labs account is preillyme [17:49:49] scfc_de: could you take look at the webgrid for the tools-info tool (even if it's not beloved by you ;) ) [17:49:52] sockets disabled, connection limit reached [17:50:29] after restart plain html and python works, but php doesn't [17:51:33] * hedonil doesn't want to rewrite the stuff in python [18:01:12] preilly: What group do you mean? [18:01:19] hedonil: Looking. [18:01:32] scfc_de: thx [18:04:25] scfc_de: I was looking for the group that Tim has https://www.mediawiki.org/wiki/HHVM these tests in [18:08:12] preilly: I think overall that's in the beta/deployment-prep project. hashar? [18:08:22] Not there. [18:08:49] hedonil: If you do "qstat -j 287331", do you see on which host the webserver is supposed to be running? [18:09:08] (I. e., tools-webgrid-01 or tools-webgrid-02.) [18:11:23] scfc_de: aiaics no hostnam fro the grid https://tools.wmflabs.org/paste/view/9ec098c9 [18:11:33] scfc_de: I think that’s right [18:12:47] scfc_de: maybe an old orphaned fcgi process ? [18:14:32] hedonil: I "webservice restart", and after that, php-cgi processes from Apr09 remained. I "kill -HUP"ed those, and now it appears functional? [18:15:30] scfc_de: yep.works now. thanks. [18:18:52] preilly: Then in principle, you need to ask any of the "Admins" in the info box at https://wikitech.wikimedia.org/wiki/Nova_Resource:Deployment-prep. I think hashar is the principal MC there, though others may have enough knowledge as well. [18:19:15] hello peter :) [18:19:49] Speaking of the devil ... :-) [18:20:14] chrismcmahonbrb: looking into the login issue [18:20:39] hashar: can you grant me access to deployment-prep on labs I’d like to help Tim with the HHVM work [18:20:39] !log tools tools-webgrid-01, tools-webgrid-02: "kill -HUP" all php-cgis that are not (grand-)children of lighttpd processes [18:20:41] Logged the message, Master [18:20:52] hashar: my labs account is preillyme [18:21:16] preilly: sync with ori as well. He did a bunch of the hhvm integration on beta [18:22:08] preilly: yo [18:22:34] tim hasn't been working on beta, he's been doing all of his work locally afaik [18:22:51] ori: oh I was talking to him yesterday and he said it was on labs [18:23:03] ori: where have you been working? [18:23:33] i've been working on labs, but i haven't been working on hhvm code, just the provisioning of it [18:23:57] there are currently two somewhat independent efforts that still need to converge: [18:23:59] I was at Facebook HQ yesterday for the Hack launch event and it was pretty exciting [18:24:12] preilly: did you see tim added a --no-fun option to the test runner? [18:26:02] preilly: anyways, there's (a) working on puppet / apache / misc. server configs in preparation for deployment [18:26:12] that's been happening on labs, but tim is not super involved with that [18:26:52] then there's (b) porting extensions to hhvm. the biggest item there is luasandbox, and tim decided to use the zend compat layer [18:27:10] but then he found it was broken in all sorts of ways so he contributed some very substantial patches [18:27:32] ori: yeah 14 patches so far [18:27:43] ori: facebook was super happy [18:27:44] including porting tsrm to hhvm and converting the entire compat layer to HNI [18:28:01] ori: yeah I read through all of his commits [18:28:14] ori: and talked with sara about them as well [18:28:36] ori: so I guess I should just ask for my own group on labs to do some extension testing? [18:29:56] preilly: we should have another project anyhow for this work, could simply call it hhvm [18:30:10] i can create one and add you if you like [18:31:14] hi ori thanks [18:31:21] wondering whether anomie tim and chad have a hhvm project on labs already [18:31:33] could be [18:31:41] ori: that would be great [18:32:15] got mistaken by the hhvm-build instance which is used to compile hhvm. That is in the integration project [18:33:30] preilly: for extensions we are porting Wikimedia PHP extensions to hhvm. Some progress is listed at https://www.mediawiki.org/wiki/Wikimedia_MediaWiki_Core_Team/Quarterly_review,_April_2014#HipHop_VM_Deployment [18:33:50] Tim / Anomie / AaronSchulz would have more details for sure [18:34:01] hashar: okay sounds good [18:34:05] ori and I are merely working on integrating it on the cluster [18:34:10] think about it as pre-ops work [18:34:20] yeah makes sense [18:34:43] since you are mostly a dev you want to sync with them. Then your work will yield us package which we can integrate on beta [18:35:07] dev (whatever instance) -> integration (beta cluster) -> ops (prod) [18:35:11] hashar: okay sounds good [18:35:23] and Chad Horohoe loves Hack already [18:35:36] hashar: yeah it’s pretty great for sure [18:35:46] from the feature list, I noticed the strong typing which is a very nice addition [18:36:38] hashar: yeah it’s really full of good stuff like async for example [18:37:39] take a look at: http://docs.hhvm.com/manual/en/hack.async.php [18:38:10] I’m totally jazzed about: async function gen_foo(int $a): Awaitable {…. [18:38:25] ori: have you tried Hack at all? [18:40:50] preilly: that almost looks like C# :p [18:41:24] valhallasw: yeah it’s funny that you say that… they’ve hired quite a few people from MSFT that worked on the C# compiler team [18:42:27] Ryan_Lane1: you’re number 1 [18:44:25] hashar: so it looks like AaronSchulz is just using his own VM but he also pointed me towards http://en.wikipedia.beta-hhvm.wmflabs.org/wiki/Special:SpecialPages [18:46:50] preilly: that is pointing to the beta cluster [18:46:56] being integrated mainly by ori [18:47:28] it is merely to work on the puppet stuff and how to get hhvm packaged for our production needs [18:47:46] so we run out of some upstream releases, nothing from some HEAD nor experimental code [18:48:59] okay makes sense can I get my own project to create a VM in for testing? [18:49:13] or rally aaron project so you can work together maybe ? :) [18:49:21] (meanwhile I dont know how to create a project hehe) [18:49:27] ha ha [18:49:42] would be faster to have Aaron to add you to its project and have him create you an instance :D [18:49:56] Okay I asked Aaron [18:50:17] I’ll also ask Eloquence I guess [18:51:03] when he logs in to IRC again [18:57:35] Anyone around that knows database shizzlness for wikipedia? [19:02:37] Damianz: mediawiki.org knows :p [19:02:58] Damianz: what's the issue? [19:02:59] ori: could you ping me when login is back on beta labs? [19:03:02] ;p [19:03:43] User table - why would users not have a registration time (or in fact all null fields) I see 490766 registered users with no registration time [19:04:11] Damianz: I think that was only introduced after a while [19:04:18] it might have also been scrubbed? [19:04:19] (idk) [19:04:34] chrismcmahon: yes, will do [19:05:39] Damianz: was introduced with wm 1.6 = 2009-ish [19:06:00] Well I have a user who's first edit was 2010 with no reg time... so that would fit. [19:06:20] Is there some sort of hacky time calculation for registration? Or should I just do what I do for anons and take now() [19:06:47] Damianz: you could find the first edit [19:06:53] but that might be heavy query [19:07:13] I'd need to find that very quickly... which probably isn't really doable [19:07:45] Damianz: for a single user, select * from revision order by rev_timestamp asc limit 1; might just work [19:09:42] use revision_userindex and "where rev_user_text={username}" [19:09:45] Seems select rev_timestamp from revision where rev_user = 'Paradoxs' order by rev_timestamp asc limit 0,1; is pretty much fast enough [19:09:55] Sadly I don't have explain rights to check indexes though [19:10:18] Nettrom: use user_id it runs much faster [19:10:46] Betacommand: better, yes, but you'll still need to use the revision_userindex table, "revision" doesn't have an index on rev_user nor rev_user_text [19:11:05] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Tables_for_revision_or_logging_queries_involving_user_names_and_IDs [19:11:20] Nettrom: there should be one on rev for user contribs [19:14:02] there's also a thread on labs-l about estimating registration dates, where halfak describes his approach and later points to a dataset: http://lists.wikimedia.org/pipermail/labs-l/2014-February/002122.html [19:18:14] valhallasw: trying to run tor on tools-login? :) [19:20:11] Nettrom: Just wanted to reference that. Always best to copy others' work :-). [19:22:36] scfc_de: took me a few minutes to find the thread, didn't remember correctly where I've seen it, was happy I found it :) [19:25:44] yuvipanda: :P just not pressing ctrl-D often enough to end up at my own box [19:26:02] valhallasw: hehe [19:26:16] !python del [19:26:16] Successfully removed python [19:26:47] yuvipanda: don't you love the python hate? :-p [19:26:57] valhallasw: and the rest of it, yea :) [19:27:00] !php [19:27:00] a beautiful language next to C. [19:27:05] !php del [19:27:05] Successfully removed php [19:27:18] !javascript [19:27:25] ok that's not too bad [19:27:30] can we also kill the 'ping' response? [19:27:35] !ping [19:27:35] !pong [19:27:38] no, not that one [19:27:55] yuvipanda: ping [19:27:56] valhallasw: hmm? [19:28:03] valhallasw: pong? [19:28:15] wm-bot also sometimes complains about people saying ': ping!' [19:28:32] not sure when though. *shrug* [19:29:57] Perhaps only with petan? [19:30:06] !log pywikibot pywikibugs: if checking the project fails, always notify #pywikipediabot and mention the exception [19:30:07] pywikibot is not a valid project. [19:30:08] what [19:30:19] Hrm, that still hasn't been deployed. That's disappointing. [19:30:23] !log local-pywikibot pywikibugs: if checking the project fails, always notify #pywikipediabot and mention the exception [19:30:25] Logged the message, Master [19:30:58] New email: labslogbot changed a page on your watchlist! [19:30:59] *grin* [19:54:52] hashar: how often does the betalabs update script run again? [19:55:29] every six minutes for code [19:55:32] and hourly for databases [19:55:39] you can guess by looking at https://integration.wikimedia.org/dashboard/ [19:55:45] and then follow each link [19:55:50] the job history would tell you yuvipanda :] [19:56:03] OK, where the hell does wmf store their interwiki table? [19:59:43] Betacommand: in localsettings.php, basically [20:00:19] why the hell dont they use the db? [20:00:48] why the hell don't you calm down? [20:01:52] valhallasw: Betacommand doesn't do calm :p [20:03:26] valhallasw: it gets really annoying when trying to port something to a local wiki when the WMF uses non-standard configs [20:03:53] Betacommand: afaik language link specifications always reside in config files [20:05:14] valhallasw: should be in interwiki table [20:05:40] https://www.mediawiki.org/wiki/Manual:Interwiki_table [20:06:49] Betacommand: Hrm. In my memory they were configured in config files, but I must be mistaken. [20:09:38] * a930913 updates valhallasw's RAM. [20:09:49] valhallasw: thats how the WMF does it [20:09:54] Betacommand: but googling for 'wmf interwiki configuration' would have brought you to https://www.mediawiki.org/wiki/Interwiki_cache [20:10:21] combine that with https://noc.wikimedia.org/conf/ , which lists a file called interwiki.cdb [20:10:50] valhallasw: its still annoying as hell to debug an issue when the wmf decides to use something no where close to the default [20:11:24] yeah, dirty commie bastards! [20:11:27] Betacommand: LOL. [20:11:34] should just make an extra SQL query every time. [20:11:40] LIKE THE REST OF US PLEBS! [20:11:47] Betacommand: because, you know, the WMF runs sites that clearly have 'default' load. [20:11:57] * yuvipanda turns off memcache everywhere [20:12:39] Betacommand: and, to make matters worse, this has actually been documented! [20:12:49] valhallasw: when debugging an issue its a pain in the ass [20:12:54] It's not like it's OpenSSL or git style documentation either ;-) [20:14:17] Betacommand: what kind of issue would depend on labs having an interwiki table or not? [20:15:52] valhallasw: I was trying to copy a few of the iw rows for my local server [20:36:36] scfc_de, Coren: Another transient failure occured. [20:36:47] With the grid queue [21:13:44] !log deployment-prep Disabled https://integration.wikimedia.org/ci/job/beta-code-update-eqiad/ to work on scap setup [21:13:46] Logged the message, Master [21:34:34] on ee-flow-extra labs, I'm back to permission problems. `sudo su - vagrant` `touch /vagrant/mediawiki/extensions/Flow/junk.test` gets permission denied [21:34:51] but vagrant owns the directory. [21:35:46] !log deployment-prep Running scap on deployment-bastion for the first time in eqiad [21:35:48] Logged the message, Master [21:36:07] yet `touch /vagrant/mediawiki/junk.test` works, same ownership [21:36:33] spagewmf: Vagrant owns /vagrant/mediawiki/extensions/Flow/? What's the permissions mask there? [21:36:58] bd808: drwxr-xr-x 12 vagrant www-data 4096 Apr 3 03:11 /vagrant/mediawiki/extensions/Flow [21:37:31] * bd808 is puzzled [21:37:42] same as drwxr-xr-x 15 vagrant www-data 4096 Apr 10 21:35 /vagrant/mediawiki , which vagrant *can* write to [21:38:23] hm, /tmp/crontab.iXO7i1: No space left on device [21:39:34] bd808: aha! `stat` shows two different vagrant users! [21:39:53] Oh! A local one and the ldap user [21:39:58] on tools-login [21:41:02] spagewmf: Nuke the local user; make sure you chown /vagrant to the ldap user's uid and reboot [21:41:25] .../Flow is owned by 601/vagrant /home/mediawiki is owned by the 1001/vagrant [21:41:55] OK [21:42:05] 601 is the uid from ldap [21:42:25] uid=601(vagrant) gid=601(vagrant) groups=601(vagrant) [21:52:20] bd808: worked great, I didn't even reboot. New Highlander slogan "There kinda should be only one" :) [21:52:57] It's better to burn out than to fade away! [22:22:21] bd808, I filed bug 63793 about the second vagrant user. I can't reproduce because project editor-engagement reached our instance limit (10/10) [22:24:43] spagewmf: I think we fixed that when physkerwelt was setting up his instances for math [22:25:17] bd808: good. I think I created ee-flow-extra around March 24th [22:25:39] hmmm… it should have been fixed before then I think [22:26:02] I'll look into it … eventually :) [22:26:40] Yup. well, it's there in case anyone else runs into it. Should I file a bug to increase editor-engagement's instance quota ? [22:28:37] * bd808 shrugs [22:29:50] spagewmf: It's not fixed. :( Easy patch though [22:30:11] Or… maybe this was the fix [22:35:33] * bd808 notes that S assigns him more bugs that his boss does :) [22:37:23] sorry! I see ori gets assigned all vagrant bugs by default, and panic. I'll file one more to add a labs-vagrant component to MediaWiki-Vagrant [22:37:52] :) [22:37:55] Should it be a component of labs? :P [22:38:16] It's a component of yuvipanda's imagination [22:38:28] It's totally cool… when it works [22:38:58] * bd808 now wants to see a yuvipanda component in bugzilla [22:39:35] * Error: yuvipanda is on the wrong side to the globe [22:39:49] *of the [23:58:00] tools.giftbot@tools-login:~$ crontab -e [23:58:00] /tmp/crontab.FEMr5X: No space left on device [23:58:01] Creation of temporary crontab file failed - aborting [23:58:08] Coren: ^ [23:59:58]